Sun Microsystems, Inc.

Oracle America, Inc.

Gregory R Ruetsch

Ilya A Sharapov

Neelakanth M Nadgir

Jerome M Verbeke

Vu Trang

Michael J Vernik

During operation, the system of an embodiment of the present invention receives a representation of multiple objective functions as well as a representation of a domain of interest for the multiple objective functions. Next, the system stores the representations in a memory within the computer system. The system then performs an interval optimization process to compute guaranteed bounds on a Pareto front for the multiple objective functions. During this interval optimization process, the system maintains influence information for boxes in the domain of interest, wherein for a given box, the influence information identifies other boxes which are in the range of influence of the given box and/or in the domain of influence of the given box. The system uses this influence information to identify boxes to be tested against each other for domination. After the boxes are tested, the system eliminates boxes which are certainly dominated by other boxes.

Using interval techniques of direct comparison and differential formulation to solve a multi-objective optimization problem

A system that solves a parametric multi-objective optimization problem in a combined design space and parameter space using interval techniques is described. The design space contains design-space variables fixed for a selected design; the parameter space contains variable parameters for the selected design. Multiple-objective functions are specified for optimization. The system initializes a design-variable box spanning the design space and performs interval optimization process on the parameter space by subdividing the design-variable box into design-variable sub-boxes, and iteratively: (1) determining parametric Pareto fronts for a design-variable sub-box using an interval optimization technique; (2) comparing parametric Pareto fronts associated with a set of design-variable sub-boxes and determining the parametric Pareto fronts certainly dominated by other parametric Pareto fronts; (3) eliminating the design-variable sub-boxes associated with the certainly dominated Pareto fronts; and (4) subdividing remaining design-variable sub-boxes. An optimized solution is produced from the remaining design-variable sub-boxes and the associated parametric Pareto fronts.

Using interval techniques to solve a parametric multi-objective optimization problem

One embodiment of the present invention provides a system that uses a computer to evaluate a function within a domain using an interval computing technique. During operation, the system receives the function and the domain over which the function is to be evaluated. Next, the system creates a set of sampling subdomains within the domain by generating a set of boxes of different sizes which are located at different positions across the domain. The system then performs evaluations on the set of sampling subdomains to generate evaluation results. Next, the system sorts the evaluation results based on box size. The system then selects a box size from the different box sizes by identifying a largest box size which achieves tight bounds on the evaluation results. The system next discretizes the domain into subdomains of the selected box size. Finally, the system uses the interval computing technique to evaluate the function on each of the subdomains wherein the interval computing technique recursively subdivides and evaluates the function on the subdomains.

Method and apparatus for initializing interval computations through subdomain sampling

System and method for the unique naming of resources in networked environments including peer-to-peer network environments. Embodiments may be used to generate network resource identifiers that uniquely identify resources such as processes over time and across the networked environment. One embodiment may be implemented as a resource name generator that generates unique identifiers for resources in networked environments. One embodiment may generate unique network resource identifiers for processes from information including, but not limited to, a host address that uniquely identifies on the network the host machine of the resource for which the unique network resource identifier is being generated, a current time (e.g. obtained from the host machine or alternatively from an external source such as an atomic time server), and a process identifier of the process that uniquely identifies the process among current processes on the host machine.

System and method for unique naming of resources in networked environments

System and method for submitting and performing computational tasks in a distributed heterogeneous networked environment. Embodiments may allow tasks to be submitted and run in parallel on a network of heterogeneous computers implementing a variety of operating environments. In one embodiment, a user on an originating node may advertise code on the network. Peer nodes that respond to the advertisement may receive the code. A job to be executed by the code may be split into separate tasks to distributed to the peer nodes that received the code. These tasks may be advertised on the network. Tasks may be assigned to peer nodes that respond to the task advertisements. The peer nodes may then work on the assigned tasks. Once a peer node's work on a task is completed, the peer node may return the results of the task to the originating node.

System and method for submitting and performing computational tasks in a distributed heterogeneous networked environment

One embodiment of the present invention provides a system that uses interval techniques to solve a multi-objective optimization problem. During operation, the system receives a representation of multiple objective functions (&#x192;, . . . , &#x192;) at a computer system, wherein (&#x192;, . . . , &#x192;) are scalar functions of a vector x=(x, . . . , x). The system also receives a representation of a domain of interest for the multiple objective functions. Next, the system performs an interval optimization process to compute guaranteed bounds on a Pareto front for the objective functions (&#x192;, . . . , f), wherein for each point on the Pareto front, an improvement in one objective function cannot be made without adversely affecting at least one other objective function. While performing the interval optimization process, the system applies a direct-comparison technique between subdomains of the domain of interest to eliminate subdomains that are certainly dominated by other subdomains.

Method and apparatus for using interval techniques to solve a multi-objective optimization problem

One embodiment of the present invention provides a system and method for determining a cache optimized ordering of cells in an unstructured graph. Cells bounding a region defined along a portion of a stored logically-defined grid are identified and each cell is added into an element of a level set array block in order of traversal through the region along the boundary. The level set array block is reordered for each additional cell in the data object that is contiguous to at least one such cell added previously to the level set array block. Each such additional cell is added into an element of the level set array block. Each cell remaining in the data object independent of any element in the level set array block is iteratively added.

Growing community of inventors

West Linn, OR, United States of America

Gregory R Ruetsch